Content uploaded by Tomasz Fraczyk
Author content
All content in this area was uploaded by Tomasz Fraczyk on Nov 02, 2020
Content may be subject to copyright.
Human Annexins A1, A2, and A8 as Potential Molecular Targets for
Ni(II) Ions
Nina E. Wezynfeld, Karolina Bossak, Wojciech Goch, Arkadiusz Bonna, Wojciech Bal,
and Tomasz Frączyk*
Institute of Biochemistry and Biophysics, Polish Academy of Sciences, Pawińskiego 5a, 02-106 Warsaw, Poland
*
SSupporting Information
ABSTRACT: Nickel is harmful for humans, but molecular
mechanisms of its toxicity are far from being fully elucidated.
One of such mechanisms may be associated with the Ni(II)-
dependent peptide bond hydrolysis, which occurs before Ser/Thr
in Ser/Thr-Xaa-His sequences. Human annexins A1, A2, and A8,
proteins modulating the immune system, contain several such
sequences. To test if these proteins are potential molecular targets
for nickel toxicity we characterized the binding of Ni(II) ions and
hydrolysis of peptides Ac-KALTGHLEE-am (A1−1), Ac-
TKYSKHDMN-am (A1−2), and Ac-GVGTRHKAL-am (A1−3),
from annexin A1, Ac-KMSTVHEIL-am (A2−1) and Ac-SAL-
SGHLET-am (A2−2), from annexin A2, and Ac-VKSSSHFNP-am
(A8−1), from annexin A8, using UV−vis and circular dichroism
(CD) spectroscopies, potentiometry, isothermal titration calorimetry, high-performance liquid chromatography (HPLC), and
electrospray ionization mass spectrometry (ESI-MS). We found that at physiological conditions (pH 7.4 and 37 °C) peptides
A1−2, A1−3, A8−1, and to some extent A2−2 bind Ni(II) ions sufficiently strongly in 4N complexes and are hydrolyzed at
sufficiently high rates to justify the notion that these annexins can undergo nickel hydrolysis in vivo. These results are discussed in
the context of specific biochemical interactions of respective proteins. Our results also expand the knowledge about Ni(II)
binding to histidine peptides by determination of thermodynamic parameters of this process and spectroscopic characterization
of 3N complexes. Altogether, our results indicate that human annexins A1, A2, and A8 are potential molecular targets for nickel
toxicity and help design appropriate cellular studies.
■INTRODUCTION
Nickel is toxic for humans, causing allergy, cancers of the
respiratory system, and other serious health problems.
1−5
Despite this, nickel is a component of stainless steel and other
alloys, which are found in coins, mobile phones, accessories,
jewelry, and many implantable materials. These objects and
materials remain in contact with human tissues and release
nickel.
6−13
Furthermore, significant amounts of this metal are
delivered into lungs from standard as well as electronic
cigarettes and from airborne dust.
14−16
Nickel, as presumably
the most frequent contact allergen, is responsible for substantial
health problems, as more than 5% of the general population of
many developed countries (e.g., Germany and Denmark) is
sensitized to this metal.
3,6
The prevalence of nickel allergy is
especially high (17%) among women.
17
Although many
hypotheses have been proposed for the mechanisms of
development of metal-induced diseases, such as nickel allergy,
there is a high possibility that many relevant mechanisms are
yet to be unveiled. One of such possibility may be related to
hydrolysis of a peptide bond catalyzed by Ni(II) ions.
The Ni(II)-dependent hydrolysis of peptide bond occurs in
amino acid sequences Yaa-Ser/Thr-Xaa-His-Zaa (where Yaa
and Zaa stand for any amino acid residue and Xaa means any
amino acid residue except Pro). The hydrolyzed bond is located
between the Yaa and Ser/Thr residues. The binding of Ni(II)
occurs through the formation of a four nitrogen (4N) square
planar complex (by the imidazole and amide of histidine, and
two preceding amides), which causes the bending of the
peptide chain, and places the nucleophilic Ser/Thr hydroxyl
group in a proximity of the peptide bond preceding this residue.
As a result, the acyl shift takes place with the formation of an
intermediate ester product, which subsequently undergoes
spontaneous hydrolysis in aqueous solution.
18−20
Sequences susceptible to Ni(II)-dependent hydrolysis are
present in many proteins, but only some of them can undergo
hydrolysis under physiological conditions. We demonstrated
that the type of amino acid residue close to the potential
hydrolysis site has a significant influence on the hydrolysis rate.
This effect is especially relevant for those amino acids that
neighbor the histidine residue, where bulky and hydrophobic
residues are preferred for the fast reaction.
18,19,21
The formation
of the initial 4N Ni(II) complex is prerequisite for the
hydrolysis to occur; thus, the accessibility of the potential
Received: August 19, 2014
Published: October 7, 2014
Article
pubs.acs.org/crt
© 2014 American Chemical Society 1996 dx.doi.org/10.1021/tx500337w |Chem. Res. Toxicol. 2014, 27, 1996−2009
cleavage site for Ni2+ ions is required. The protein sequences
harboring the cleavage sites should therefore be exposed on the
protein surface and sufficiently flexible to adopt the square
planar structure of the complex.
22
However, local bends of the
protein main chain may predispose it to such conformation,
resulting in the hydrolysis reaction faster than expected on the
basis of sequence alone.
23
Annexins are proteins that bind phospholipid membranes in
a calcium-dependent manner. This binding is performed by
their conserved C-terminal domain. The N-terminal domain
interacts with other proteins, depending on the annexin type.
The specificity of these interactions has an impact on roles
played by different annexins.
24
Functions of annexin A1 include
(1) regulation of the innate and adaptive immune systems,
25
(2) participation in the plasma membrane repair system,
26
(3)
helicase activity,
27
and more. Annexin A2 participates in
immunological processes.
28
It is worth mentioning that the
expression of annexin A2 was found to be induced by nickel in
human HaCaT keratinocyte line.
29
Annexin A8 is a poorly
characterized member of the annexin family. It is known to
bind F-actin (similarly as annexin A1 and A2
30
) and is
associated with late endosomes.
31,32
To find out whether human annexins A1, A2, and A8 are
potential targets for the toxicity of Ni(II) ions we synthesized
peptides that represent fragments of these proteins potentially
susceptible to Ni(II)-dependent hydrolysis. We also synthe-
sized a reference peptide Ac-GGASRHWKF-am, with an amino
acid sequence corresponding to that of the positive hydrolysis
control sequence established in our previous studies. Molecular
modeling, UV−vis and circular dichroism (CD) spectroscopies,
potentiometry, isothermal titration calorimetry, high-perform-
ance liquid chromatography (HPLC), and electrospray
ionization mass spectrometry (ESI-MS) were used to character-
ize Ni(II) binding to these peptides and Ni(II)-dependent
peptide bond hydrolysis.
■EXPERIMENTAL SECTION
Materials. N-α-9-Fluorenylmethyloxycarbonyl (Fmoc) amino acids
were purchased from Novabiochem (Merck). Trifluoroacetic acid
(TFA), piperidine, O-(benzotriazol-1-yl)-N,N,N′,N′-tetramethyluro-
nium hexafluorophosphate (HBTU), triisopropylsilane (TIS), and
N,N-diisopropylethylamine (DIEA) were purchased from Merck.
Acetic anhydride was purchased from Sigma-Aldrich. TentaGel S
RAM resin was obtained from Rapp Polymer GmbH.
Molecular Modeling. Crystallographic structures of annexins A1,
A2, and A8 (PDB codes: 1MCX, 2HYU, and 1W45, respectively) were
taken to simulations performed in Discovery Studio 4.0 Visualizer.
Ni(II) ion with a square planar geometry was incorporated to
respective potential sites. Conformations of complexes were obtained
by geometry optimization utilizing Dreiding-like force field.
33
Peptide Synthesis. All peptides were synthesized in the solid
phase according to the Fmoc protocol
34
using an automatic peptide
synthesizer (Prelude, Protein Technology). The syntheses were
accomplished on a TentaGel S RAM resin (RAPP Polymere
GmbH), using HBTU as a coupling reagent, in the presence of
DIEA. Both acetylation of the N-terminus and cleavage were done
manually. The acetylation was carried out in 10% acetic anhydride in
DCM. The cleavage was done by the cleavage mixture composed of
95% TFA, 2.5% TIS, and 2.5% water. Crude peptides were isolated
from cleavage mixtures by precipitation by the addition of cold diethyl
ether. Following precipitation, peptides were dissolved in water and
lyophilized. Finally, they were purified by HPLC, and their identities
were checked by ESI-MS, as described before.
19
UV−vis and Circular Dichroism Spectroscopies. The UV−vis
spectra were recorded on the LAMBDA 950 UV/vis/NIR
spectrophotometer (PerkinElmer) over the spectral range 330−850
nm. The CD spectra were recorded on the J-815 CD spectrometer
(JASCO) over the spectral range 300−650 nm. For both methods,
path length was 1 cm, and samples containing 0.95 mM peptide and
0.9 mM Ni(NO3)2were titrated with small portions of concentrated
NaOH in the pH range 3.0−11.5, at 25 °C.
Potentiometry. Potentiometric titrations were performed on a
907 Titrando Automatic Titrator (Metrohm), using a Biotrode
combined glass electrode (Metrohm), calibrated daily by nitric acid
titrations.
35
One hundred millimolar NaOH (carbon dioxide free) was
used as a titrant. Samples (1.5 mL) were prepared by dissolving
peptides in 4 mM HNO3/96 mM KNO3to obtain 0.8−1.5 mM
peptide concentrations. The Ni(II) complex formation was studied
using samples in which the molar ratios of peptide to Ni(II) were
1:0.9, 1:0.45, and 1:0.3. The pH range for all potentiometric titrations
was 2.7−11.6. All experiments were performed under argon at 25 °C.
Three titrations were included simultaneously into calculations for
protonation, and five for Ni(II) complexation. The data were analyzed
using the SUPERQUAD and HYPERQUAD programs.
36,37
Standard
deviations provided by these software and reported here have
statistical nature and do not include potential systematic errors.
Isothermal Titration Calorimetry. Calorimetric titrations were
carried out on the Nano ITC Standard Volume calorimeter (TA
Instruments). The sample cell (950 μL) was filled with a peptide
solution, and the reference cell was filled with Milli-Q water. The
syringe (250 μL) was loaded with a Ni(II) solution. Milli-Q water was
degassed under vacuum for 15 min before sample preparation.
Furthermore, peptide and Ni(II) solutions were degassed for 5 min
before loading into the cell and the syringe. The peptide solutions
contained 2 mM peptide, 100 mM H3BO3, and 64 mM KNO3at pH
9.0. The ionic strength was 0.1 M. The Ni(II) solutions contained 20
mM Ni(NO3)2and 64 mM KNO3. Typically, 2 μL of the Ni(II)
solution was added to the peptide solution at 1200−1500 s intervals
using a stirring speed of 250 rpm. To prevent the metal-dependent
hydrolysis of peptide bond during the experiments, the measurements
were performed at 5 °C. The blank linear functions were calculated on
the basis of the last measurement points, where the observed heat flow
resulted almost exclusively from the dilution of Ni(II) solution. The
blank corrections were made by subtracting values of the blank linear
function from the raw data. The data were analyzed by the
NanoAnalyze Software (TA Instruments).
HPLC Measurements of Hydrolysis Rates. Samples containing
0.8 mM peptide, 1 mM Ni(NO3)2, and 20 mM HEPES at pH 8.2 or
7.4 were incubated at 25, 37, 45, or 60 °C. Twenty microliter aliquots
were periodically collected from the sample and acidified by addition
of 20 μL of 2% TFA to break down hydrolytic Ni(II) complexes in
order to stop the hydrolysis reaction. The products of hydrolysis were
separated by HPLC and their identity checked on ESI-MS.
Kinetics Analysis. To determine rate constants for the hydrolysis
reaction we used the iterative optimization algorithm based on the
Levenberg−Marquardt method, which was adapted to minimize the
squared distance between the experimental data and theoretical curves
obtained from numerical solutions of the corresponding system of the
following equations (eqs 1−3):
=− + −
S
t
R kSt St M S
d
d()( () )
4N 1 0 0 (1)
=− + − −
IP
t
R kSt St M S kIPt
d
d()( () ) ()
4N 1 0 0 2 (2)
=
P
t
kIPt
d
d()
2(3)
They were used to calculate the first (k1) and the second rate
constant (k2) on the basis of changes in time of substrate (S),
intermediate product (IP), and products (P) concentrations depend-
ing on the ratio of 4N hydrolytic complexes to the rest of substrate
species (R4N) at a given pH, and the initial concentration of the
substrate (S0) and metal ions (M0). All calculations were performed in
the Wolfram Mathematica 8 environment.
Chemical Research in Toxicology Article
dx.doi.org/10.1021/tx500337w |Chem. Res. Toxicol. 2014, 27, 1996−20091997
The full description of the model used to describe the kinetics of the
metal dependent hydrolysis, including approximations that could be
used when the concentration of 4N complexes are in excess or in
deficiency to the rest of the substrate species throughout the
experiment, is available in the Supporting Information.
■RESULTS
Analysis of 3D Structures of Annexins A1, A2, and A8.
Structures of human annexins A1, A2, and A8 are available in
the RCSB PDB database. This allowed us to analyze geometries
of potential sites of Ni(II) binding and Ni(II)-dependent
peptide bond hydrolysis. However, for human annexin A1, the
crystallographic data are available only for Cαatoms of the
main chain of the protein (PDB code: 1AIN). The amino acid
sequence alignment of porcine and human annexin A1 shows
high similarity (89% of identical amino acid residues), with full
identity in the regions with potential sites of Ni(II) binding and
hydrolysis. Structural alignment of the human (PDB code:
1AIN) and porcine (PDB code: 1MCX) proteins confirms high
similarity. Therefore, we used the structure of porcine annexin
A1 (PDB code: 1MCX), which includes bound calcium ions.
For annexin A2 we chose the crystal structure of human protein
complexed with calcium ions and heparin fragments (PDB
code: 2HYU). These heparin fragments do not affect the
conformation of potential Ni(II) binding and hydrolysis sites
(as compared to the respective structure without heparin; PDB
code: 2HYW). For annexin A8 we chose the structure of the
human protein without calcium ions (PDB code: 1W45), as
these ions do not influence the site of potential Ni(II) binding
and hydrolysis (as compared to the respective structure with
calcium ions; PDB code: 1W3W). Moreover, the 1W45
structure contains a longer fragment of the N-terminus of the
protein where the potential Ni(II) binding and hydrolysis site is
located.
To test the possibility of formation of 4N square planar
Ni(II) complexes in the above-mentioned annexins we
simulated the Ni(II) binding with the use of geometry
optimization utilizing Dreiding-like force field
33
in Discovery
Studio 4.0 Visualizer. The fragments of proteins containing sites
for Ni(II) binding and hydrolysis, together with corresponding
simulated complexes are shown in Figure 1. All these sites are
located in coils, assuring the flexibility of the polypeptide chain.
Accordingly, all simulated complexes were acquired without
disturbing large parts of proteins, which means that the process
of Ni(II) binding would not need to overcome excessive energy
barriers. This analysis shows that the sites are accessible for
solutes such as metal ions and can adopt conformations
conforming to square planar complexes. These features can
promote the binding of Ni(II) ions to potential hydrolytic sites
in annexins.
In the amino acid sequence of human annexin A1 there are
three Yaa-Ser/Thr-Xaa-His-Zaa sites that may potentially bind
Ni(II) and undergo Ni(II)-dependent peptide bond hydrolysis.
They are 101Thr-Gly-His103,244Ser-Lys-His246, and 291Thr-Arg-
His293 (Figure 1A−F). They are located on the convex side of
the molecule. This side binds Ca2+ ions and interacts with
phospholipids. The following Ca2+ binding residues are located
in the proximity of potential Ni(II) sites: Lys97, Leu100,
Glu105, Asp253, Leu256, Gly288, Met286, and Gly290.
Although it is not observed in simulated structures, this
proximity suggests that the Ni(II) binding can affect Ca2+
binding and vice versa. Interestingly, imidazole groups of His103
and His293 residues are 4 Å apart and interact via π−π
stacking.
Human annexin A2 exists in two isoforms resulting from
alternative splicing. The second isoform differs from the
canonical one by the longer N-terminus (18 residues). Here, we
will use the numbering for the canonical sequence. It contains
two Ni(II) hydrolysis sites. The N-terminal 3Thr-Val-His5site
is not resolved in crystal structures, indicating that it is flexible
or assumes multiple conformations. Indeed, NMR and CD
spectroscopic characterization of the N-terminal fragment of
annexin A2 showed that, although it adopts an α-helical
conformation in a membrane-mimetic environment, it has
random structure in aqueous solution.
38
The 92Ser-Gly-His94
site is located on the surface of the convex side of molecule
(Figure 1G,H). Adjacent amino acid residues Lys88, Leu91,
and Glu96 bind Ca2+ ions. Furthermore, His94 interacts with a
fragment of heparin. Thus, the Ni(II) binding is likely to affect
interactions of annexin A2 with Ca2+ ions and heparin.
The potential Ni(II) binding sequence in human annexin A8,
18Ser-Ser-His20 (Figure 1I,J) is located on the concave side of
the protein, opposite to the Ca2+ binding side. The fragment is
on the surface of the protein in a coiled, flexible structure.
The above analysis shows that all Yaa-Ser/Thr-Xaa-His-Zaa
sites in human annexins A1, A2, and A8 are easily accessible for
Figure 1. Fragments of human annexins A1, A2, and A8 with potential
Ni(II) binding sites. Shown are native proteins (A, C, E, G, I) and with
Ni(II) ion bound (B, D, F, H, J) as modeled with the use of geometry
optimization utilizing Dreiding-like force field
33
in Discovery Studio
4.0 Visualizer. Shown are fragments with 101TGH103 (A,B), 244SKH246
(C,D), and 291TRH293 (E,F) from annexin A1, 92SGH94 (G,H) from
annexin A2, and 18SSH20 (I,J) from annexin A8. Amino acid residues
expected to compose the Ni(II) complex are shown in a stick
representation. The sequences studied in this article are violet; Ca,
green; Ni2+, yellow; heparin, white sticks.
Chemical Research in Toxicology Article
dx.doi.org/10.1021/tx500337w |Chem. Res. Toxicol. 2014, 27, 1996−20091998
Ni(II) ions. Therefore, we decided to characterize the Ni(II)
binding and susceptibility to Ni(II)-dependent peptide bond
hydrolysis of peptides with sequences taken from the annexins.
Sequences of these peptides in the protein context and labels
used for them in the text are presented in Table 1.
Ni(II) Complexation. We used UV−vis and circular
dichroism spectroscopies, potentiometry, and isothermal
titration calorimetry to characterize the Ni(II) binding to
annexin-derived peptides. Table 2 presents cumulative
logarithmic protonation constants of all peptides calculated
on the basis of potentiometric titrations. All peptides contain a
single His residue, with pKvalues in the range of 6.1 to 6.8.
Most of the peptides also contain a Lys or a Tyr residue, with
pK9.3−10.6, and acidic residues, with pK3.6−4.8.
As potentiometry records only H+concentration changes
upon the metal ion binding, rather than metal specific
properties, we utilized UV−vis results to select the correct
stoichiometric model of Ni(II) binding and to verify Ni(II)
complex formation constants calculated on the basis of
potentiometric titrations. The distributions of Ni(II) species
described by particular models were compared to the pH
dependence of absorbance at the maximum of d−d bands of
low-spin complexes. We used root-mean-square deviation
(rmsd) as a parameter to measure differences between UV−
vis spectroscopic parameters and species distributions based on
these models. Ni(II) can be coordinated by imidazole ring of
histidine residue (1N complex), by the imidazole and two
amide nitrogens (3N complex), or by the imidazole and three
amide nitrogens (4N complex). Previous research showed that
2N complexes are not observed in hydrolytic peptides
containing Yaa-Ser/Thr-Xaa-His-Zaa sequences and that 3N
complexes are formed from 1N species through cooperative
deprotonation of two amide nitrogens.
18
We examined at least
two models for each peptide. In the first model we assumed the
formation of 1N, 3N, and 4N complexes. In the second model
we disregarded 3N complexes. The simulated amounts of 3N
complexes, whenever considered, were always low compared to
corresponding 4N complexes, and never exceeded 25% of all
Ni(II) species. We were able to confirm the existence of 3N
complexes in UV−vis spectra only for A1−3, A2−1, A2−2, and
A8−1 peptides. The comparisons of Ni(II) binding models
considered are presented in Figure S1 of the Supporting
Information. The verified Ni(II) binding constants are
presented in Table 3. In Figure 2 we present examples of
Ni(II) species distributions compared to the absorbance at the
maximum of respective d−d bands. At pH above 4, Ni(II)
binds to peptides via a His imidazole ring nitrogen (1N
complex). Its presence could not be confirmed for A2−1,
probably as a result of a lower concentration of this peptide in
potentiometric experiments, due to its low solubility. In general,
1N complexes did not exceed 20% of all Ni(II) species, which
could also significantly hinder the detection of this complex for
the A2−1 peptide. As expected, the UV−vis spectra at pH of
the maximum of molar fraction for 1N complexes did not differ
significantly from those at pH below 4, which confirms that 1N
complexes are high-spin (octahedral), similarly to the Ni(II)
aqua ion.
The good fit of the absorbance at the d−d band maximum
with the sum of 3N and 4N complexes helped us conclude that
3N complexes of A1−3 and A8−1 peptides contain low-spin
Ni(II). On the contrary, for A2−1 and A2−2 peptides, the
absorbance of Ni(II) complexes at the d−d band maxima
corresponds to the sum of 4N complexes rather than the sum
of 3N and 4N complexes. Thus, Ni(II) in 3N complexes of
A2−1 and A2−2 is high-spin.
Low-spin 4N complexes are hydrolytically active species.
Their Ni(II) complex formation constants are highlighted in
bold in Table 3. It is worth noting that almost every studied
peptide can form more than one 4N complex, due to
deprotonations of Lys or Tyr residues. The only exception,
Table 1. Sequences of the Studied Peptides
human
protein label amino acid sequence of
peptide position of the histidine
residue in the protein
a
annexin
A1 A1−1 Ac-KALTGHLEE-am 103
A1−2 Ac-TKYSKHDMN-am 246
A1−3 Ac-GVGTRHKAL-am 293
annexin
A2 A2−1 Ac-KMSTVHEIL-am 5
b
/23
c
A2−2 Ac-SALSGHLET-am 94
b
/112
c
annexin
A8 A8−1 Ac-VKSSSHFNP-am 20
a
The numbering from protein sequences including the initiator
methionine is shown.
b
The positions are valid for isoform 1, chosen as
canonical sequence from the two produced by alternative splicing.
c
The positions are from isoform 2.
Table 2. Logarithmic Protonation Constants for Annexin
Peptides Determined at 25 °C and I= 0.1 M (KNO3)
a
peptide HL H2LH
3LH
4LH
5L
A1−1 10.18(1) 16.97(1) 21.78(1) 25.68(1)
A1−2 10.61(1) 20.99(1) 30.34(1) 36.70(1) 40.31(1)
A1−3 10.05(1) 16.13(1)
A2−1 9.85(1) 16.19(1) 20.51(1)
A2−2 6.66(1) 10.98(1)
A8−1 10.19(1) 16.56(1)
a
Standard deviations on the least significant digits, provided by
HYPERQUAD,
37
are given in parentheses.
Table 3. Logarithmic Ni(II) Complex Formation Constants for Annexin Peptides Determined at 25 °C and I= 0.1 M (KNO3)
a
peptide NiH3L NiH2L NiHL NiL NiH−1L NiH−2L NiH−3L
A1−1 n.d.
b
n.d. 12.61(2) n.d. n.d. −11.50(1) −21.59(1)
A1−2 32.43(3) n.d. n.d. 9.53(1) −0.08(1) −10.67(1) −21.44(1)
A1−3 n.d. n.d. 12.53(2) n.d. −3.13(1) −11.21(1) −21.61(1)
A2−1 n.d. n.d. n.d. n.d. −4.31(3) −12.11(1) −21.80(1)
A2−2 n.d n.d. n.d. 2.90(1) n.d. −13.33(2) −21.24(1)
A8−1 n.d. n.d. 12.55(3) n.d. −2.86(1) −10.86(1) −21.10(1)
a
Standard deviations on the least significant digits, provided by HYPERQUAD,
37
are given in parentheses.
b
n.d.; not detected. Values for 4N
complexes are bold.
Chemical Research in Toxicology Article
dx.doi.org/10.1021/tx500337w |Chem. Res. Toxicol. 2014, 27, 1996−20091999
the A2−2 peptide contains neither Lys nor Tyr residue and
forms only one 4N complex in the studied pH range.
The low-spin 3N and 4N complexes are detectable not only
in UV−vis but also in CD spectra in the region specific for d−d
transitions (350−600 nm). While UV−vis spectra of Ni(II)
complexes are very similar to each other for all studied
peptides, we observed a variety of CD spectral patterns,
especially in terms of the number of extrema (Figure 3). In the
range of 300−650 nm, we observed just one minimum and one
maximum for peptides A1−1 and A1−3. The CD spectra of
Ni(II) complexes of A2−1 and A8−1 peptides have an
additional maximum at about 570 nm. A2−2 has one maximum
at 456 nm but also two much smaller minima at 385 and 562
nm. For the A1−2 complex we noticed three extrema of
comparable intensities.
On the basis of potentiometric data, we used the least-
squares calculations to deconvolute UV−vis and CD spectra for
each low-spin complex species. Table 4 presents their molar
absorption coefficients. For the A1−2 peptide overlapping
deprotonations of two Lys and one Tyr residues yielded four
4N species, which could not be deconvoluted reliably. In the
course of preliminary calculations we detected that two
intermediate species, NiH−1L and NiH−2L, may have very
similar spectra and consequently merged them in the
calculations. Generally, the spectra of different Ni(II)
complexes for the same peptide have similar patterns (see
Figures S2 and S3 of the Supporting Information). The
exception is the NiL complex of A1−2forwhichwe
distinguished an additional minimum of a very low intensity
at short wavelengths. The wavelengths of the maximum
absorbance (λmax)ofd−d bands for a given peptide differ by
less than 4 nm for A1−1, A1−2, and A2−1 peptides, where,
according to stoichiometry of complexes, only 4N species are
low-spin. For A1−3andA8−1, where potentiometric
calculations indicated that both 3N and 4N complexes should
have low-spin character, the shift was 29 and 24 nm,
respectively. According to the ligand field theory, an additional
amide nitrogen involved in the Ni(II) binding in 4N complexes
should cause a larger difference in energy between d orbitals
than three nitrogen ligands as in 3N complexes. In other words,
bands of 4N complexes should be shifted toward shorter
wavelengths relative to the 3N complexes, as we indeed
observed for A1−3 and A8−1. This confirms that A1−3 and
A8−1 peptides form low-spin 3N complexes.
Calorimetric titrations were carried out at pH 9.0 to ensure
the high percentage of 4N complexes (>98%) and at 5 °Cto
inhibit the hydrolysis reaction. Thermodynamic parameters of
the analyzed process are presented in Table 5. The
representative titration is shown in Figure 4. We could not
obtain satisfactory results for Ni(II) binding to A2−1 due to
insufficient solubility of this peptide. Conditional Ni(II)
complex formation constant values for studied peptides are in
the range 1.1−6.8 ×104M−1. To compare binding constants
obtained by ITC and potentiometry, we calculated conditional
Ni(II) binding constants at pH 9.0 on the basis of
potentiometric results. They are 20−40 times higher than
those obtained from ITC. This discrepancy could be a result of
an interaction between Ni(II) and the buffer. Although we did
not detect such interactions in control titrations of borate with
Ni(II) ions, there is a literature report by Tilak et al.
39
indicating that borate and nickel form a weak Ni(H2BO3)2
complex.
Defining the association constant of the Ni(II) borate (B)
complex (Kbuff)byeq4,
=KMB
MB
[]
[][]
buff 2
2(4)
one can easily obtain a formula converting the apparent binding
constant of the Ni(II) complexation to the studied peptide
obtained from ITC (Kapp) into the conditional binding constant
at pH 9 (Kc,eq5)as
=+×KK K B(1 [ ] )
capp buff 2 (5)
Figure 2. Species distributions of Ni(II) complexes of selected peptides: Ac-KALTGHLEE-am (A1−1; left), Ac-KMSTVHEIL-am (A2−1; in the
middle), and Ac-VKSSSHFNP-am (A8−1; right), at 25 °C, calculated for concentrations used in UV−vis and CD titrations (0.95 mM peptides and
0.9 mM Ni(II)) using stability constants from Tables 2 and 3. The common scale left-side axes represent Ni(II) molar fractions. Ni(II) species are
color-marked as follows: Ni(II) aqua ion, black; 1N complex, red; 3N complex, blue; 4N complexes, green and orange. Pink dotted lines show the
sum of 4N complexes, while navy dashed lines show the sum of 3N and 4N complexes. The variable scale right-side axes provides values of
absorbance and ellipticity at the d−d band maximum of low-spin complexes: UV−vis absorbance, black circles; CD ellipticity, blue, pink, and violet
triangles. Absorbance and ellipticity values are omitted for clarity. Species distributions for other peptides are available in Figure S1 of the Supporting
Information.
Chemical Research in Toxicology Article
dx.doi.org/10.1021/tx500337w |Chem. Res. Toxicol. 2014, 27, 1996−20092000
We followed the assumption of Tilak et al. that only H2BO3
−
ions bind Ni(II).
39
Taking that pKof H3BO3at 5 °C is 9.4,
40
the log Kbuff explaining the discrepancy between ITC and
potentiometric results should be in the range of 4.4−4.7. This is
in a good agreement with results of Tilak et al. (log K= 3.8−
4.9).
39
Therefore, we can safely use ITC results to compare
Ni(II) affinities of the studied peptides. In agreement with
potentiometric results, A1−2 and A8−1 peptides have the
highest affinity to Ni(II) ions, followed by A2−2 and A1−3
peptides. A1−1 has the lowest affinity for Ni(II) ions, with the
value of binding constant more than six times lower than for
A1−2 peptide.
The observed enthalpy changes (ΔHobs) range between 27.6
and 47.7 kJ ×mol−1(Table 5). They have positive values,
indicating that Ni(II) coordination is an endothermic reaction.
The highest ΔHwas noticed for A8−1, the lowest for A1−1.
The value of the heat absorbed after addition of Ni(II) ions to
peptide solution is a result of Ni(II) binding to peptides and
the sum of energy of all other processes occurring after
injection, including dilution and buffer protonation. The heat of
dilution was subtracted. The buffer protonation was the result
of proton release from peptides upon Ni(II) binding. At pH 9.0
almost all histidine residues are deprotonated spontaneously, so
Ni(II) binding only requires deprotonation of three amide
nitrogens. According to literature data,
40
the enthalpy of borate
protonation at 5 °C is about −18.6 kJ ×mol−1. Assuming that
released protons react only with the buffer, the contribution of
this process is about −55.8 kJ ×mol−1. These considerations
yield estimated values of Ni(II) binding enthalpy changes
(ΔHNi(II) bind) ranging between 83.4 and 103.5 kJ ×mol−1
(Table 5). We did not include the possible protonation of
lysine or tyrosine residues, as we calculated that it has a minor
buffering effect in experimental conditions because of a much
higher concentration of buffer (100 mM) from that of the
peptide (2 mM).
The observed entropy changes (ΔSobs) range between 176.4
J×mol−1×K−1for A1−1 and 260.2 J ×mol−1×K−1for A8−1
(Table 5). The correction of enthalpy change values by the
buffer protonation effect led to a corresponding increase in the
entropy changes associated with Ni(II) binding (ΔSNi(II) bind)
by about 200 J ×mol−1×K−1. It suggests that Ni(II) binding is
the entropy-favored reaction under ITC experiment conditions.
It is mainly the result of water release from nickel aqua ions
after Ni(II) binding to peptide.
Hydrolysis of Peptides. All studied peptides undergo
Ni(II)-dependent peptide bond hydrolysis. Ni(II) ions
selectively cleave a peptide bond preceding serine or threonine
residue in Yaa-Ser/Thr-Xaa-His-Zaa sequences. Separation of
Figure 3. CD spectra of low-spin Ni(II) complexes of peptides: Ac-KALTGHLEE-am (A1−1), Ac-TKYSKHDMN-am (A1−2), Ac-GVGTRHKAL-
am (A1−3), Ac-KMSTVHEIL-am (A2−1), Ac-SALSGHLET-am (A2−2), and Ac-VKSSSHFNP-am (A8−1), at different pH values, coded with
rainbow colors from red (the lowest pH ∼3) to dark blue (the highest pH 11.5). The spectra of apopeptides, exhibiting no extrema in this spectral
range, are marked by dotted gray lines (mostly not visible as they have zero value throughout all the spectra). The titrations were carried out at 25
°C.
Chemical Research in Toxicology Article
dx.doi.org/10.1021/tx500337w |Chem. Res. Toxicol. 2014, 27, 1996−20092001
reactants by HPLC allowed us to detect four species in all
cases: a substrate, an intermediate product (IP), and N- and C-
terminal final products (see Figures S4 and S5 of the
Supporting Information). The identity of these species was
confirmed by ESI-MS, but because of the same mass of the
substrate and the IP, a distinction between them required an
additional comparison of their differing retention times.
We performed the peptide hydrolysis reactions at pH 7.4 and
8.2. The former value was chosen to refer to physiological
conditions; the latter, to easily compare our results with those
Table 4. Parameters of d−d Bands of Low-Spin Square-Planar Ni(II) Complexes of Annexin Peptides
UV−vis CD
peptide Ni(II) species λmax (nm) ε(dm3×mol−1×cm−1)λext (nm) Δε(dm3×mol−1×cm−1)
A1−1 NiH−2L 4N 455 135(1) 449 1.70(4)
519 −0.69(1)
NiH−3L 4N 454 140(2) 448 1.79(6)
513 −0.69(3)
A1−2 NiL 4N 457 100(1) 391 −0.11(3)
443 0.57(3)
497 −1.23(8)
559 0.74(5)
NiH−1L/NiH−2L
a
4N 457 102(1) 438 0.54(7)
496 −1.2(2)
559 0.5(1)
NiH−3L 4N 455 100(2) 436 0.60(7)
495 −1.4(2)
562 0.6(1)
A1−3 NiH−1L 3N 457 129(7) 442 0.72(4)
501 −1.2(3)
NiH−2L 4N 432 119(2) 422 1.15(3)
493 −1.49(3)
NiH−3L 4N 428 120(2) 423 1.11(4)
494 −1.56(5)
A2−1 NiH−2L 4N 452 93(3) 433 0.7(1)
493 −1.0(2)
568 0.06(3)
NiH−3L 4N 448 103(3) 430 1.1(1)
492 −1.6(2)
568 0.15(3)
A2−2 NiH−3L 4N 454 121(1) 385 −0.05(5)
456 1.6(2)
562 −0.07(6)
A8−1 NiH−1L 3N 464 103(5) 448 0.9(1)
506 −1.0(1)
569 0.18(5)
NiH−2L 4N 448 105(2) 428 0.74(2)
498 −1.22(3)
574 0.10(2)
NiH−3L 4N 440 104(2) 419 0.84(2)
492 −1.32(3)
573 0.08(5)
a
Since in the result of preliminary calculations we detected that, for A1−2, two intermediate species, NiH−1L and NiH−2L, may have very similar
spectra, we merged them in the calculations.
Table 5. Thermodynamic Parameters for Ni(II) Binding to Annexin Peptides Obtained from ITC Experiments in 100 mM
H3BO3,I= 0.1 M, at pH 9.0 and 5 °C
peptide Kaobs
(M−1×104)
a
Katheor
(M−1×104)
b
ΔGobs
(kJ ×mol−1)
a
ΔHobs
(kJ ×mol−1)
a
ΔHNi(II) bind
(kJ ×mol−1)
c
ΔSobs
(J ×mol−1×K−1)
a
ΔSNi(II) bind
(J ×mol−1×K−1)
c
A1−1 1.1(1) 21.4 −21.5 27.6(6) 83.4 176.4 377.1
A1−2 6.8(8) 131.8 −25.7 32.1(3) 87.9 208.1 408.4
A1−3 1.8(2) 58.9 −22.5 39.7(5) 95.5 223.5 424.2
A2−2 1.6(2) 63.1 −22.4 33.7(7) 89.5 202.0 402.3
A8−1 4.3(3) 100.0 −24.7 47.7(5) 103.5 260.2 460.9
a
The values obtained directly from ITC experiments.
b
The values calculated on the basis of potentiometric results.
c
The values calculated with
consideration of the protonation of buffer.
Chemical Research in Toxicology Article
dx.doi.org/10.1021/tx500337w |Chem. Res. Toxicol. 2014, 27, 1996−20092002
published earlier by us for other amino acid sequences. The
amount of 4N complexes was low at pH 7.4, as calculated using
potentiometric and spectroscopic results. Under conditions of
hydrolysis experiments, the concentration of 4N complexes did
not exceed 2% of the total peptide concentration. Furthermore,
the product of hydrolysis binds Ni(II) ions much more tightly
than the substrate,
18
scavenging Ni(II) from the system, so the
concentrations of Ni(II) aqua ions and 4N Ni(II) substrate
complexes decreased systematically during the experiment.
These conditions have a major impact on equations describing
hydrolysis used for the analysis of results of our experiments.
The rate constant equations used previously are valid for higher
pH values, where the degree of complexation is much
higher.
18,19
In our experimental conditions the hydrolytic 4N
complex is not only consumed by the reaction, but its
concentration is additionally decreased by a shift of complex-
ation equilibrium accompanying the effective dilution of the
reacting system in the course of reaction. In order to
accommodate this effect, we developed new equations (given
in the Supporting Information), which use potentiometry-
derived species distribution data and directly yield the
maximum at the specific temperature, pH-independent rate
constant for the first step of the reaction (k1Ind), corresponding
to a 100% formation of the hydrolytic 4N complex. The value
corresponding to the first pH-dependent constant, which we
used and presented previously, can be easily obtained by
multiplying the first pH independent constant (k1Ind) by the
molar fraction of the hydrolytic 4N complex at the beginning of
the reaction. From such calculation one can obtain the values of
rate constants observed at specific pH, e.g., 7.4 (k17.4) or 8.2
(k18.2).
The second step of the reaction, which is the hydrolysis of
the intermediate product described by the k2value, is
independent from the Ni(II) species distribution, although it
is also pH-dependent. The pH dependences of the first and the
second step of the metal-dependent hydrolysis are different.
For the first step of reaction, it is associated with the formation
of 4N complexes; for the second one, with the concentration of
hydroxide ions, which can catalyze hydrolysis of an ester bond
in the intermediate product.
41
Thus, in the case of the second
rate constant (k2) we calculated values valid for specific pH.
Because of the low amount of 4N complexes at pH 7.4 (e.g.,
for A1−1 about 0.16%), we decided to present rate constants
calculated on the basis of experiments performed at pH 8.2
because they provide smaller errors of determination of the
amount of 4N complexes. Rate constants, k1Ind,k18.2, and k28.2,
for the hydrolysis at pH 8.2 are presented in Table 6.
At 37 °C, the highest k1Ind and k18.2 values were for A1−3,
followed by A8−1. In the group of peptides with the highest
k28.2 values, we found A8−1 and A1−2. The lowest k1Ind,k18.2,
and k28.2 constants were noted for A2−1 and A1−1.
We used the linear Arrhenius plot to describe how
temperature influences the rate constants k18.2 and k28.2 (Figure
5). The values of k18.2 and k28.2 (Figure 5), as well as k1Ind,k17.4,
and k27.4 (Figure S6 of the Supporting Information), fulfill the
Arrhenius equation for all studied peptides. Analyzing the
temperature dependence of individual rate constants, we
observe no significant differences in slopes between the
peptides. However, the values of k1are more dependent on
temperature than k2.
The hydrolysis of the ester-containing IP is the rate limiting
step of the peptide cleavage for all studied peptides at the
experimental conditions. The amount of IP accumulating in the
initial phase of all hydrolysis experiments is a combined effect
of the rates of the first and the second step of the reaction.
Specifically, the higher the k1in relation to k2, the higher the
amount of IP observed. The highest amount of IP was detected
for A1−3 (e.g., 85% of the total peptide at pH 8.2 and 37 °C;
Figure S7 of the Supporting Information), as its k18.2 was 17.5
times higher than k28.2. The lowest amount of IP was observed
Figure 4. ITC titration of 2 mM Ac-TKYSKHDMN-am peptide (A1−
2), 100 mM H3BO3, 64 mM KNO3, pH 9.0 with 20 mM Ni(NO3)2
and 64 mM KNO3at 5 °C. The upper plot shows raw data from the
experiment; the lower plot shows the absorbed heat (per mol of
injectant) in each injection (black dots), with the fitting of the model
with assumed 1:1 interaction stoichiometry (green line).
Table 6. First Order Rate Constants Determined for Ni(II)-Dependent Hydrolysis of the Studied Peptides Calculated on the
Basis of Experiments Performed at pH 8.2
a
k(s−1×10−5)
25 °C37°C45°C
peptide k1Ind k18.2 k28.2 k1Ind k18.2 k28.2 k1Ind k18.2 k28.2
A1−1 0.321(1) 0.081(1) 0.031(1) 1.941(6) 0.492(2) 0.126(1) 6.27(2) 1.589(5) 0.294(1)
A1−2 1.148(2) 0.648(2) 2.47(2) 6.75(2) 3.811(8) 8.67(4) 19.86(2) 11.22(2) 20.32(4)
A1−3 21.18(9) 5.910(3) 0.353(1) 100.0(6) 27.9(2) 1.581(6) 480(5) 134(2) 3.86(2)
A2−1 0.944(2) 0.169(1) 0.145(1) 6.04(2) 1.082(4) 0.455(2) 15.6(1) 2.789(2) 0.968(4)
A2−2 1.014(5) 0.289(1) 0.547(3) 6.06(2) 1.878(6) 1.770(6) 16.79(5) 4.781(2) 3.467(9)
A8−1 11.38(5) 3.891(2) 2.523(9) 56.6(2) 19.37(7) 10.33(3) 139.3(7) 47.65(3) 22.08(8)
a
k1Ind, maximum at specific temperature and pH-independent rate constant for the 1st step of the Ni(II)-dependent hydrolysis; k18.2, the rate
constant for the 1st step of the Ni(II)-dependent hydrolysis at pH 8.2; k28.2, the rate constant for the 2nd step of the Ni(II)-dependent hydrolysis at
pH 8.2.
Chemical Research in Toxicology Article
dx.doi.org/10.1021/tx500337w |Chem. Res. Toxicol. 2014, 27, 1996−20092003
for A1−2 (e.g., 25% of the total peptide at pH 8.2 and 37 °C),
as its k18.2 was even lower than k28.2.Ask1decreased more
rapidly than k2at pH 7.4, when compared to pH 8.2, the
amounts of observed IP were smaller, ranging from 7% of the
total A1−2 to 42% of the total A1−3, at 37 °C. We also
detected a linear temperature dependence of the maximum
amount of IP (except A1−3 at pH 8.2; Figure S7 of the
Supporting Information), which results from the above-
mentioned dependences of inverse temperature (T−1) on the
ln(k1) and ln(k2).
We also determined the half-times of substrate decay (t1/2S)
and the half-times of formation of final products (t1/2P) (Table
7) because both of these processes could be important from the
toxicological point of view. At pH 7.4 and 37 °Ct1/2S and t1/2P
were the shortest for A1−3 and A8−1 and the longest for A2−
1 and A1−1.
Increasing pH to 8.2 does not change the order of peptides
in terms of susceptibility to Ni(II)-dependent hydrolysis.
Nevertheless, we observed a significant reduction of t1/2S at
pH 8.2 compared to t1/2S at pH 7.4, from about 40 times for
A1−1 to about 82 times for A2−2. The t1/2P was shortened
from about 6 times for A1−1 up to 24 times for A2−2 peptide.
At pH 8.2, differences between t1/2S and t1/2P were bigger when
compared to the values observed at pH 7.4.
The hydrolysis of the reference peptide (Ac-GGASRHWKF-
am) was the fastest among all of the studied peptides. At pH
7.4, t1/2S and t1/2P were 2.2 times shorter as compared to
annexin peptides characterized by the shortest t1/2S and t1/2P.At
pH 8.2, these differences were not so big, and t1/2S of the
reference peptide was similar to t1/2S of A1−3orA8−1. The
t1/2P for the reference peptide was comparable to t1/2P for A8−1
(Table 7).
■DISCUSSION
Characteristics of Complexes. To characterize Ni(II)
complexes, we used methods applied successfully in our
previous research on Ni(II) complexes of peptides susceptible
to Ni(II)-dependent hydrolysis: potentiometry and UV−vis
and CD spectroscopies. We supplemented our research
methodology with ITC as an auxiliary method for determi-
nation of binding constants and thermodynamic parameters.
According to our best knowledge, systematic studies of
kinetics of formation of various Ni(II) complexes of His-n(n>
3, where ndenotes position in peptide) and His-3 peptides
were not performed. Nevertheless, it seems clear that His-n
complexes are formed much faster than His-3 complexes. For
example, slow reactivity of the order of minutes to hours is
observed frequently for the latter, often precluding potentio-
metric determination of their stabilities. Such behavior was not
seen in His-ncomplexes studied by us, as well as other research
groups.
18,41−44
In the case of annexin peptides we observed the
equilibration of complex formation reactions essentially within
the mixing time (few seconds) of the samples.
The 1N complexes are octahedral species anchored at the
imidazole nitrogen, the only nitrogen donor accessible, due to
its low pKavalue. We detected 1N, 3N, and 4N complexes.
Among 3N complexes we found high-spin species for A2−1
and A2−2, and low-spin (square planar) species for A1−3 and
A8−1, judged by the absence or presence of a relatively intense
d−d band close to 450 nm. The low-spin 3N complexes have
spectra similar to low-spin 4N; however, they are systematically
red-shifted by 24 to 29 nm. It is, according to our best
knowledge, the first clear-cut account of d−d parameters of
low-spin 3N Ni(II) complexes of peptides. We demonstrated
that 3N complexes of annexin model peptides are either low- or
high-spin. In previous studies on Ni(II) complexes of peptides
susceptible to Ni(II)-dependent hydrolysis, also a mixture of
high- and low-spin 3N complexes was observed among Ni(II)
complexes for the same peptide, e.g., for Ac-GASRHAKFL-
am.
18
The 3N complexes in His peptides are a little mysterious. In
many cases they were found to be inexistent or very minor
species. When observed in amounts sufficient to characterize
their spectroscopic properties, they were found to be high-spin,
low-spin, or a mixture of these two cases. In general, the spin
state of a Ni(II) complex depends on the relationship between
the ligand field strength of the equatorial vs potential axial
ligands (e.g., thermochromism and solvatochromism).
45−48
The gradual substitutions of water oxygens with peptide
nitrogens bring the initially octahedral complexes closer to the
spin crossover limit. In complexes involving main chain peptide
nitrogen coordination equatorial 1N and 2N complexes are
always high-spin and 4N complexes are always low-spin. The
Figure 5. Arrhenius plot of the first rate constants, k18.2 (A), and the
second rate constants, k28.2 (B), describing Ni(II)-dependent peptide
bond hydrolysis at pH 8.2 for peptides: Ac-KALTGHLEE-am (A1−1),
black squares; Ac-TKYSKHDMN-am (A1−2), red circles; Ac-
GVGTRHKAL-am (A1−3),bluetriangles;Ac-KMSTVHEIL-am
(A2−1), pink triangles; Ac-SALSGHLET-am (A2−2), green squares;
and Ac-VKSSSHFNP-am (A8−1), navy triangles.
Table 7. Half-Times of Substrate Decay (t1/2S) and
Formation of Hydrolysis Products (t1/2P) of the Studied
Peptides at 37 °C and pH 7.4 and 8.2
t1/2S (h) t1/2P (h)
peptide pH 7.4 pH 8.2 pH 7.4 pH 8.2
A1−1 1251 31 1321 231
A1−2 168 4.1 182 7.7
A1−329 0.6 109 14
A2−1 1076 16 1295 79
A2−2 610 7.5 702 24
A8−145 0.9 60 3.0
reference
a
13 0.5 28 2.2
a
The peptide Ac-GGASRHWKF-am, with an amino acid sequence
corresponding to that of the positive nickel hydrolysis control
sequence established in our previous research.
18,19
Values for the
fastest processes for annexin peptides in physiological conditions are
bold.
Chemical Research in Toxicology Article
dx.doi.org/10.1021/tx500337w |Chem. Res. Toxicol. 2014, 27, 1996−20092004
3N complexes of peptides studied in this work appear to be
near or at the boundary, and so their high-spin (octahedral) or
low-spin (square planar) character is determined by minor
alterations of electronic density of the third nitrogen in the
structure of the complex, that is the amide nitrogen of the
residue preceding the His residue. The difficulty of studying
this phenomenon stems from the fact that these 3N complexes
are usually minor species, present only at low concentrations,
due to high cooperativity in the formation of three-ring systems
of final 4N complexes, as Hill coefficient for this process was
always higher than 2 for all studied peptides. In many of the
peptides studied they were not even detected. In order to find
the rules for the formation of 3N complexes we calculated
protonation-corrected stability constants *K3N according to eq
6:
18,49,50
ββ*= −
−
K
l
og log log
xNNiH L H
L
nx n (6)
The βHnLcorresponds to the protonation of the histidine
residue and the βNiHn−xLis a Ni(II) cumulative stability constant
of complexes with xnitrogen atoms coordinating metal ion.
The protonation-corrected stability constants enable compar-
isons of complexes having different protonation stoichiome-
tries. In particular, they allowed us to compensate for the
numeric effects of deprotonations of nonbonding Tyr and Lys
residues in values of βconstants. In Figure 6 we compared log
*K3N values of high-spin, low-spin, and mixed 3N complexes of
annexin peptides studied here, of hydrolytic peptides used in
our previous research,
18
and of analogues of the model peptide
of histone H2A.
42
Generally, among 3N complexes, low-spin
species are the most, and high-spin are the least stable ones.
Complexes with intermediate log *K3N values are a mixture of
high- and low-spin species. The exception is Ac-
GASKHWKFL-am, which forms high-spin 3N complex despite
a high value of log *K3N.
18
The residue preceding His seems to
be the most important for the 3N complex spin state, and Arg
residue seems to be preferred in this position for low-spin
complexes. Noteworthy, it was demonstrated that Arg side
chain lowers the basicity of its amine nitrogen down to pKa
7.11−7.53
49,51−54
from a typical value of ∼7.8. This effect may
also have an influence on proximate amide nitrogens
coordinating the metal ion. Mlynarz et al. showed that the
lower is the basicity of an amide nitrogen, the higher is the
stability of the Ni(II) complexes.
55
Altogether, this principle
explains the highest stabilities of 3N complexes with the Arg
residue.
While we observed heterogeneity of spectroscopic properties
of 3N complexes, d−d band parameters of 4N complexes for all
studied peptides are very similar to each other and consistent
with the results for two species of 4N complex (NiH−2L and
NiH−3L) of the peptide derived from histone H2A and its
analogues
42,43
and the peptide derived from the C-terminus of
histone H4.
44
The protonation-corrected stability constants of the first (at
the lowest pH) 4N complexes, *K4N, for annexin peptides,
calculated according to eq 6, revealed the highest stability of
Ni(II) complexes with A1−2, A1−3, and A8−1 peptides, with
log *K4N values −27.17, −27.34, and −27.41, respectively. The
remaining log *K4N values are −28.03 for A2−1, −28.30 for
A2−2, and −28.46 for A1−1. The range of log *K4N values
observed for annexin peptides is similar to those obtained in
our previous study on model hydrolytic peptides (from −28.12
to −27.05)
18
as well as to the Ni(II) complexes of the peptide
derived from histone H2A and its analogues (from −28.58 to
−27.26),
42,43
and model peptide of C-terminus of histone H4
(−27.92).
44
Therefore, the 4N Ni(II) complexes containing
imidazole nitrogen and three amide nitrogen donors have
rather uniform stabilities (within 1.6 log units).
When comparing patterns of CD spectra of Ni(II) complexes
of model annexin peptides, we noted that A1−1 and A1−3
peptides have the same pattern as observed previously for Ac-
GASRHWKFL-am, a model hydrolytic peptide,
56
and for
peptide Ac-GTHS-am.
57,58
The patterns of CD spectra of A2−
1 and A8−1 peptides are more similar to that observed for Ac-
TESAHK-am,
42
with an additional small positive band at longer
wavelengths. The A1−2 peptide has the same CD spectra
pattern as that for Ac-TYTEHA-am, the model peptide from C-
terminus of histone H4,
44
with more pronounced positive band
at 550 nm. CD spectra of A2−2 peptide are unusual, as they
generally have only a positive band in the measured spectral
range.
Complex Stabilities and Hydrolysis in the Context of
Biology. The nickel-dependent peptide bond hydrolysis was
observed in viable cultures of CHO (Chinese hamster ovary),
NRK-52 (rat renal tubular epithelium), and HPL1D (human
lung epithelium) cells treated with Ni(II) ions for one to seven
days.
59
The cleaved bond was found to be in histone H2A
between Glu and Ser residues in sequence ESHHK. It should
be noted that the analytical method used in the cited study
allowed us to observe only the histones. The above-mentioned
value of −28.58 for the protonation-corrected logarithmic
stability constant (log *K4N) for the Ni(II) complex with Ac-
TESHHK-am (a model sequence from histone H2A) was
reported.
42
The comparison of this value with the respective
values for peptides studied here shows that annexin peptides
form more stable Ni(II) complexes, compared to the peptide
derived from H2A. Furthermore, we calculated that the
amounts of 4N hydrolytic complexes for Ac-TESHHK-am
peptide at pH 7.4 and for concentrations used in our hydrolysis
experiments are significantly (from 5.5 to 65 times, Figure S8 of
the Supporting Information) lower than those observed for
Figure 6. Protonation-corrected logarithmic stability constants of
Ni(II) 3N complexes, log *K3N, and spin states of these complexes.
The considered peptides are annexin peptides marked by green
squares, Ac-GVGTRHKAL-am (TRHK); Ac-KMSTVHEIL-am
(TVHE); Ac-SALSGHLET-am (SGHL); Ac-VKSSSHFNP-am
(SSHF); and other, previously studied peptides
18,42
marked by black
squares, Ac-GASGHAKFL-am (SGHA); Ac-GASAHWKFL-am
(SAHW); Ac-GASKHWKFL-am (SKHW); Ac-GASRHAKFL-am
(SRHA); Ac-GASRHWKFL-am (SRHW); Ac-GATRHWKFL-am
(TRHW); Ac-TESAHK-am (SAHK); Ac-TESHAK-am (ESHA).
Chemical Research in Toxicology Article
dx.doi.org/10.1021/tx500337w |Chem. Res. Toxicol. 2014, 27, 1996−20092005
model peptides of annexins presented here. Since Ni(II) was
shown to bind and cleave H2A in cultured cells, it seems to be
likely that the same processes can also occur for annexin A1,
A2, or A8.
Peptides from histone H2B were thoroughly tested
considering binding of Ni2+ ions.
60−63
It was suggested that
formation of such complexes may change conformation of the
whole protein inducing epigenetic and promutagenic effects.
However, taking into account relatively weak binding at
physiological pH, the significance of these findings remains to
be established.
Toll-like receptor 4 (TLR4) was proposed by Schmidt et
al.
64
as a target molecule for Ni2+ ions in nickel allergy. They
stated that a Ni2+ ion bridges two monomers of TLR4 leading
to its direct activation. However, the relevance of that finding in
the context of nickel allergy was called into question in the
recent paper of Vennegaard et al.,
65
who found that
epicutaneous exposure (resembling common challenge of the
human skin by nickel from, e.g., jewelry) of mice to nickel can
induce allergy independently of TLR4, in contrast to
intradermal injections of nickel performed by Schmidt et al.
64
This example shows that the mechanisms of nickel allergy are
far from being fully elucidated. There is a need for further
research to reveal the complete spectrum of pathophysiology of
Ni(II).
Rates of Hydrolysis. From the physiological point of view,
it is more intuitive to compare the amounts of reaction
products at pH 7.4 and 37 °C, after, e.g., 24 h (Figure 7),
instead of using values of rate constants. In these conditions
more than one-third of the peptide A1−3 with the sequence
including His293 of annexin A1 is converted to the ester-
containing intermediate product. At the same time, one-fifth of
the peptide A8−1 with the sequence from annexin A8
underwent a complete two-step hydrolysis reaction. In the
case of peptides from annexin A2, the largest amount of IP
(3%) was observed for A2−2. These results suggest that
annexins A1 and A8, and to some extent also A2, may be
toxicological targets for Ni(II) ions in the context of peptide
bond hydrolysis.
Potential Impact of Ni(II) Binding and Hydrolysis on
Properties of Proteins. The binding of Ni(II) or subsequent
hydrolysis can have several different effects; it can (1) change
the conformation of the protein and disturb its native function,
(2) influence the stability of the protein, and (3) create
epitopes recognized by the immune system as non-native. As
mentioned above, annexins regulate the immune system. Thus,
malfunctioning of the immune system observed, e.g., in allergy
can be induced by any of these three modes of action.
Annexin A1 can inhibit phospholipase A2, cyclooxygenase-2,
and inducible nitric oxide synthase, enzymes involved in
inflammatory processes.
66
Furthermore, one of the potential
Ni(II) binding sites (244SKH246) in annexin A1 is part of
246HDMNKVLDL254, a sequence that was found to inhibit
antigen-induced proliferation of T cells.
67
Interestingly, the
same site is close to 254LELKGD259 sequence, which is known
as a nuclear export signal.
68
Finally, all three potential Ni(II)
binding sites are in the proximity of Ca2+ binding residues.
Thus, one can assume that binding of Ni(II) and subsequent
change of the conformation, being a result either of a binding or
an acyl shift, as well as hydrolysis of peptide bond in annexin
A1 may have an influence on important physiological processes.
The N-terminal part of annexin A2, containing potential
Ni(II) binding residues 3TVH5, consists of a nuclear export
sequence: 4VHEILCKLSL13.
69
Furthermore, residues Val4, Ile7,
Leu8, and Leu11 participate in the interaction with the
S100A10 protein.
69
There is a redox active Cys9 residue in
the same region. This cysteine residue is reversibly oxidized by
H2O2and reduced by the thioredoxin system; thus, annexin A2
is a probable antioxidant molecule.
70
Annexin A2 accumulates
in the nucleus in response to oxidative stress, and this is
probably one of the means of DNA protection. In the absence
of oxidative stress, annexin A2 is exported from the nucleus.
70
In this context, it is important to mention that Ni(II)
complexes with peptides were found to induce DNA breakage
in the presence of H2O2.
53,71−73
Thus, binding of Ni(II) to
3TVH5may have an influence on these processes. The Ni(II)-
dependent cleavage of peptide bond between Ser2 and Thr3
could also transform isoform 2 of annexin A2 with the longer
N-terminal part to the analogue of isoform 1. As mentioned
above, another potential Ni(II) binding site, 92SGH94, is close
to Ca2+ binding residues. In the result, Ni(II) binding may also
impair calcium-dependent functions of annexin A2.
The binding of Ni(II) and peptide bond hydrolysis in
annexin A8 is probably most relevant for smokers and other
humans exposed to high nickel level in the air, as the highest
levels of this protein was observed in human lung
endothelium.
74
The results presented above permit us to consider the
toxicological impact of three chemical processes differing in
their velocity: (1) Ni(II) binding, (2) formation of the
intermediate product, and (3) formation of the final products.
Chelation of nickel by amino acid sequences including the His
residue in position further than 3 is fast (in the range of
seconds, as mentioned above), and its duration is correlated
with a short contact of human skin with, e.g., coins. It has to be
emphasized that even such short contact is sufficient to elicit
allergic reaction.
11
However, formation of IP and final products
is slower. It seems to correlate with the fact that the nickel
allergy is usually linked with type IV hypersensitivity, which is
known to develop days after the contact with allergen.
75
Perspective. We demonstrated that all peptides from
human annexins studied here undergo Ni(II)-dependent
hydrolysis. Analysis of 3D structures of annexins A1, A2, and
A8 showed that potential hydrolytic sites are accessible for
Ni(II) ions and can adopt conformations required for the
square planar complex formation in native proteins. The next
Figure 7. Molar fraction of an intermediate product (blue) and
hydrolysis final products (red) after 24 h at pH 7.4 and 37 °C, for
peptides Ac-KALTGHLEE-am (A1−1), Ac-TKYSKHDMN-am (A1−
2), Ac-GVGTRHKAL-am (A1−3), Ac-KMSTVHEIL-am (A2−1), Ac-
SALSGHLET-am (A2−2), and Ac-VKSSSHFNP-am (A8−1).
Chemical Research in Toxicology Article
dx.doi.org/10.1021/tx500337w |Chem. Res. Toxicol. 2014, 27, 1996−20092006
step of our research is to react whole proteins with Ni(II) ions,
followed by experiments on cell lines. These studies will allow
us to verify possible toxicological effects of Ni(II)-dependent
hydrolysis on the structure and functions of annexins.
■ASSOCIATED CONTENT
*
SSupporting Information
Detailed description of kinetics of the metal-dependent
hydrolysis; species distributions of Ni(II) complexes for
considered binding models, including rmsd results; deconvo-
luted UV−vis and CD spectra of low-spin Ni(II) complexes;
representative HPLC chromatograms and the time dependence
of relative amount of peptide species during hydrolysis;
Arrhenius plots for hydrolysis performed at pH 7.4; a
temperature dependence of the maximum amount of IP; a
comparison of 4N complexes amounts at pH 7.4. This material
is available free of charge via the Internet at http://pubs.acs.org.
■AUTHOR INFORMATION
Corresponding Author
*Tel: +48-22-592-5766. Fax: +48-22-659-4636. E-mail:
tfraczyk@ibb.waw.pl.
Funding
This study was partially financed by Polish National Science
Centre, Grant No. DEC-2011/01/D/NZ1/03501. This work
was also supported in part by the project “Metal dependent
peptide hydrolysis. Tools and mechanisms for biotechnology,
toxicology and supramolecular chemistry”, carried out as part of
the Foundation for Polish Science TEAM/2009-4/1 program,
cofinanced from European Regional Development Fund
resources within the framework of Operational Program
Innovative Economy. The equipment used was sponsored in
part by the Centre for Preclinical Research and Technology
(CePT), a project cosponsored by European Regional
Development Fund and Innovative Economy, The National
Cohesion Strategy of Poland.
Notes
The authors declare no competing financial interest.
■ABBREVIATIONS
DCM, dichloromethane; DIEA, N,N-diisopropylethylamine;
HBTU, O-(benzotriazol-1-yl)-N,N,N′,N′-tetramethyluronium
hexafluorophosphate; ITC, isothermal titration calorimetry;
IP, intermediate product; P, products; S, substrate; TIS,
triisopropylsilane; TFA, trifluoroacetic acid; TLR4, Toll-like
receptor 4; Xaa, any amino acid residue except proline; Yaa,
Zaa, any amino acid residue
■REFERENCES
(1) Beyersmann, D., and Hartwig, A. (2008) Carcinogenic metal
compounds: recent insight into molecular and cellular mechanisms.
Arch. Toxicol. 82, 493−512.
(2) Bal, W., Protas, A. M., and Kasprzak, K. S. (2011) Genotoxicity of
metal ions: chemical insights. Met. Ions Life Sci. 8, 319−373.
(3) Garg, S., Thyssen, J. P., Uter, W., Schunch, A., Johansen, J. D.,
Menne, T., Belloni Fortina, A., Statham, B., and Gawkrodger, D. J.
(2013) Nickel allergy following European Union regulation in
Denmark, Germany, Italy and the U.K. Br. J. Dermatol. 169, 854−858.
(4) Goodman, J. E., Prueitt, R. L., Thakali, S., and Oller, A. R. (2011)
The nickel ion bioavailability of the carcinogenic potential of nickel-
containing substances in the lung. Crit. Rev. Toxicol. 41, 142−174.
(5) Yao, Y., and Costa, M. (2014) Toxicogenomic effect of nickel and
beyond. Arch. Toxicol. 88, 1645−1650.
(6) Schnuch, A., Wolter, J., Geier, J., and Uter, W. (2011) Nickel
allergy is still frequent in young German females, probably because of
insufficient protection from nickel-releasing objects. Contact Dermatitis
64, 142−150.
(7) Darlenski, R., Kazandjieva, J., and Pramatarov, K. (2012) The
many faces of nickel allergy. Int. J. Dermatol. 51, 523−530.
(8) Hamann, C. R., Hamann, D., Hamann, C., Thyssen, J. P., and
Liden, C. (2012) The cost of nickel allergy: a global investigation of
coin composition and nickel and cobalt release. Contact Dermatitis 68,
15−22.
(9) Thyssen, J. P., Gawkrodger, D. J., White, I. R., Julander, A.,
Menne, T., and Liden, C. (2012) Coin exposure may cause allergic
nickel dermatitis: a review. Contact Dermatitis 68,3−14.
(10) Aquino, M., Mucci, T., Chong, M., Lorton, M. D., and Fonacier,
L. (2013) Mobile phones: potential sources of nickel and cobalt
exposure for metal allergic patients. Pediatr. Allergy Immunol. Pulmonol.
26, 181−186.
(11) Julander, A., Midander, K., Herting, G., Thyssen, J. P., White, I.
R., Wallinder, I. O., and Liden, C. (2013) New UK nickel-plated steel
coins constitute an increased allergy and eczema risk. Contact
Dermatitis 68, 323−330.
(12) Svedman, C., Möller, H., Gruvberger, B., Gustavsson, C. G.,
Dahlin, J., Persson, L., and Bruze, M. (2014) Implants and contact
allergy: are sensitizing metals released as haptens from coronary
stents? Contact Dermatitis 71,92−97.
(13) Midander, K., Kettelarij, J., Julander, A., and Liden, C. (2014)
Nickel release from white gold. Contact Dermatitis 71, 108−128.
(14) Mann, E., Ranft, U., Eberwein, G., Gladtke, D., Sugiri, D.,
Behrendt, H., Ring, J., Schafer, T., Begerow, J., Wittsiepe, J., Kramer,
U., and Wilhelm, M. (2010) Does airborne nickel exposure induce
nickel sensitization? Contact Dermatitis 62, 355−362.
(15) Pappas, R. S., Fresquez, M. R., Martone, N., and Watson, C. H.
(2014) Toxic metal concentrations in mainstream smoke from
cigarettes available in the USA. J. Anal. Toxicol. 38, 204−211.
(16) Williams, M., Villarreal, A., Bozhilov, K., Lin, S., and Talbot, P.
(2013) Metal and silicate particles including nanoparticles are present
in electronic cigarette cartomizer fluid and aerosol. PLoS One 8,
e57987.
(17) Thyssen, J. P., Linneberg, A., Menne, T., and Johansen, J. D.
(2007) The epidemiology of contact allergy in the general population:
prevalence and main findings. Contact Dermatitis 57, 287−299.
(18) Kopera, E., Krężel, A., Protas, A. M., Belczyk, A., Bonna, A.,
Wysłouch-Cieszyńska, A., Poznański, J., and Bal, W. (2010) Sequence-
specific Ni(II)-dependent peptide bond hydrolysis for protein
engineering: reaction conditions and molecular mechanism. Inorg.
Chem. 49, 6636−6645.
(19) Krężel, A., Kopera, E., Protas, A. M., Poznański, J., Wysłouch-
Cieszyńska, A., and Bal, W. (2010) Sequence-specific Ni(II)-
dependent peptide bond hydrolysis for protein engineering.
Combinatorial library determination of optimal sequences. J. Am.
Chem. Soc. 132, 3355−3366.
(20) Podobas, E. I., Bonna, A., Polkowska-Nowakowska, A., and Bal,
W. (2014) Dual catalytic role of the metal ion in nickel-assisted
peptide bond hydrolysis. J. Inorg. Biochem. 136, 107−114.
(21) Protas, A. M., Ariani, H. H., Bonna, A., Polkowska-Nowakowska,
A., Poznański, J., and Bal, W. (2013) Sequence-specific Ni(II)-
dependent peptide bond hydrolysis for protein engineering: active
sequence optimization. J. Inorg. Biochem. 127,99−106.
(22) Kopera, E., Belczyk-Ciesielska, A., and Bal, W. (2012)
Application of Ni(II)-assisted peptide bond hydrolysis to non-
enzymatic tag removal. PLoS One 7, e36350.
(23) Kurowska, E., Sasin-Kurowska, J., Bonna, A., Grynberg, M.,
Poznański, J., Knizewski, L., Ginalski, K., and Bal, W. (2011) The
C2H2 zinc finger transcription factors are likely targets for Ni(II)
toxicity. Metallomics 3, 1227−1231.
(24) Gerke, V., Creutz, C. E., and Moss, S. E. (2005) Annexins:
linking Ca2+ signaling to membrane dynamics. Nat. Rev. Mol. Cell Biol.
6, 449−461.
Chemical Research in Toxicology Article
dx.doi.org/10.1021/tx500337w |Chem. Res. Toxicol. 2014, 27, 1996−20092007
(25) D’Acquisto, F., Perretti, M., and Flower, R. J. (2008) Annexin-
A1: a pivotal regulator of the innate and adaptive immune systems. Br.
J. Pharmacol. 155, 152−169.
(26) Draeger, A., Monastyrskaya, K., and Babiychuk, E. B. (2011)
Plasma membrane repair and cellular damage control: the annexin
survival kit. Biochem. Pharmacol. 81, 703−712.
(27) Hirata, A., Corcoran, G. B., and Hirata, F. (2010) Carcinogenic
heavy metals replace Ca2+ for DNA binding and annealing activities of
mono-ubiquitinated annexin A1 homodimer. Toxicol. Appl. Pharmacol.
248,45−51.
(28) Swisher, J. F., Burton, N., Bacot, S. M., Vogel, S. N., and
Feldman, G. M. (2010) Annexin A2 tetramer activates human and
murine macrophages through TLR4. Blood 115, 549−558.
(29) Acevedo, F., Serra, M. A., Ermolli, M., Clerici, L., and
Vesterberg, O. (2001) Nickel-induced proteins in human HaCaT
keratinocytes: annexin II and phosphoglycerate kinase. Toxicology 159,
33−41.
(30) Gerke, V., and Moss, S. E. (2002) Annexins: from structure to
function. Physiol. Rev. 82, 331−371.
(31) Goebeler, V., Ruhe, D., Gerke, V., and Rescher, U. (2006)
Annexin A8 displays unique phospholipid and F-actin binding
properties. FEBS Lett. 580, 2430−2434.
(32) Goebeler, V., Poeter, M., Zeuschner, D., Gerke, V., and Rescher,
U. (2008) Annexin A8 regulates late endosome organization and
function. Mol. Biol. Cell 19, 5267−5278.
(33) Hahn, M. (1995) Receptor surface models. 1. Definition and
construction. J. Med. Chem. 38, 2080−2090.
(34) Chan, W., and White, P., Eds. (2000) Fmoc Solid Phase Peptide
Synthesis: A Practical Approach, Oxford University Press, Oxford.
(35) Irving, H., Miles, M. G., and Pettit, L. D. (1967) A study of
some problems in determining the stoicheiometric proton dissociation
constants of complexes by potentiometric titrations using a glass
electrode. Anal. Chim. Acta 38, 475−488.
(36) Gans, P., Sabatini, A., and Vacca, A. (1985) SUPERQUAD: an
improved general program for computation of formation constants
from potentiometric data. J. Chem. Soc., Dalton Trans., 1195−1200.
(37) Gans, P., Sabatini, A., and Vacca, A. (1996) Investigation of
equilibria in solution. Determination of equilibrium constants with the
HYPERQUAD suite of programs. Talanta 43, 1739−1753.
(38) Hong, Y. H., Won, H. S., Ahn, H. C., and Lee, B. J. (2003)
Structural elucidation of the protein- and membrane-binding proper-
ties of the N-terminal tail domain of human annexin II. J. Biochem. 134,
427−432.
(39) Tilak, B. V., Gendron, A. S., and Mosoiu, M. A. (1977) Borate
buffer equilibria in nickel refining electrolytes. J. Appl. Electrochem. 7,
495−500.
(40) Goldberg, R. N., Kishore, N., and Lennen, R. M. (2002)
Thermodynamic quantities for the ionization reactions of buffers. J.
Phys. Chem. Ref. Data 31, 231−370.
(41) Belczyk-Ciesielska, A., Zawisza, I. A., Mital, M., Bonna, A., and
Bal, W. (2014) Sequence-specific Cu(II)-dependent peptide bond
hydrolysis: similarities and differences with the Ni(II)-dependent
reaction. Inorg. Chem. 53, 4639−4646.
(42) Mylonas, M., Krężel, A., Plakatouras, J. C., Hadjiliadis, N., and
Bal, W. (2002) The binding of Ni(II) ions to terminally blocked
hexapeptides derived from the metal binding -ESHH- motif of histone
H2A. J. Chem. Soc., Dalton Trans., 4296−4306.
(43) Krężel, A., Mylonas, M., Kopera, E., and Bal, W. (2006)
Sequence-specific Ni(II)-dependent peptide bond hydrolysis in a
peptide containing threonine and histidine residues. Acta Biochim. Pol.
53, 721−727.
(44) Karavelas, T., Malandrinos, G., Hadjiliadis, N., Mlynarz, P.,
Kozlowski, H., Barsam, M., and Butler, I. (2008) Coordination
properties of Cu(II) and Ni(II) ions towards the C-terminal peptide
fragment −TYTEHA−of histone H4. Dalton Trans., 1215−1223.
(45) Flamini, A., Fares, V., and Pifferi, A. (2000) On the spectral
similarities between 1,2,6,7-tetracyano-3,5-dihydro-3,5-diiminopyrroli-
zinato complexes and phthalocyanines: X-ray crystal and molecular
structure of two mixed monopyrrolizinato nickel(II) complexes with
the 2,4-tert-butylacetylacetonide ligand. Eur. J. Inorg. Chem., 537−544.
(46) Ohtsu, H., and Tanaka, K. (2004) Equilibrium of low- and high-
spin states of Ni(II) complexes controlled by the donor ability of the
bidentate ligands. Inorg. Chem. 43, 3024−3030.
(47) Thies, S., Bornholdt, C., Kohler, F., Sonnichsen, F. D., Nather,
C., Tuczek, F., and Herges, R. (2010) Coordination-induced spin
crossover (CISCO) through axial bonding of substituted pyridines to
nickel-porphyrins: σ-donor versus π-acceptor effects. Chem.Eur. J.
16, 10074−10083.
(48) Venkataramani, S., Jana, U., Dommaschk, M., Sonnichsen, F. D.,
Tuczek, F., and Herges, R. (2011) Magnetic bistability of molecules in
homogenous solution at room temperature. Science 331, 445−448.
(49) Bal, W., Kozlowski, H., Robbins, R., and Pettit, L. D. (1995)
Competition between the terminal amino and imidazole nitrogen
donors for coordination to Ni(II) ions in oligopeptides. Inorg. Chim.
Acta 231,7−12.
(50) Kozłowski, H., Bal, W., Dyba, M., and Kowalik-Jankowska, T.
(1999) Specific structure-stability relations in metallopeptides. Coord.
Chem. Rev. 184, 319−346.
(51) Sovago, I., Radomska, B., Schon, I., and Nyeki, O. (1990)
Copper(II) and nickel(II) complexes of diastereomeric segments of
thymopoietin. Polyhedron 9, 825−830.
(52) Pettit, L. D., Bal, W., Bataille, M., Cardon, C., Kozlowski, H.,
Leseine-Delstanche, M., Pyburn, S., and Scozzafava, A. (1991) A
thermodynamic and spectroscopic study of the complexes of the
undecapeptide Substance P, of its N-terminal fragment and of model
pentapeptides containing two prolyl residues with copper ions. J.
Chem. Soc., Dalton Trans., 1651−1656.
(53) Bal, W., Jezowska-Bojczuk, M., and Kasprzak, K. S. (1997)
Binding of nickel(II) and copper(II) to the N-terminal sequence of
human protamine HP2. Chem. Res. Toxicol. 10, 906−914.
(54) Kowalik-Jankowska, T., Rajewska, A., Szeszel-Fedorowicz, W.,
and Kopińska, D. (2005) Bonding of copper(II) ions by proctolin
analogues modified in fifth position of the peptide chain. Polyhedron
24, 443−450.
(55) Młynarz, P., Gaggelli, N., Panek, J., Stasiak, M., Valensin, G.,
Kowalik-Jankowska, T., Leplawy, M. L., Latajka, Z., and Kozłowski, H.
(2000) How the α-hydroxymethylserine residue stabilizes oligopeptide
complexes with nickel(II) and copper(II) ions. J. Chem. Soc., Dalton
Trans., 1033−1038.
(56) Ariani, H. H., Polkowska-Nowakowska, A., and Bal, W. (2013)
Effect of D-amino acid substitutions on Ni(II)-assisted peptide bond
hydrolysis. Inorg. Chem. 52, 2422−2431.
(57) Turi, I., Kallay, C., Szikszai, D., Pappalardo, G., Di Natale, G., De
Bona, P., Rizzarelli, E., and Sovago, I. (2010) Nickel(II) complexes of
the multihistidine peptide fragments of human prion protein. J. Inorg.
Biochem. 104, 885−891.
(58) Timari, S., Turi, I., Varnagy, K., and Sovago, I. (2014) Studies
on the formation of coordination isomers in the copper(II) and
nickel(II) complexes of peptides containing histidyl residues.
Polyhedron 79,72−79.
(59) Karaczyn, A. A., Bal, W., North, S. L., Bare, R. M., Hoang, V. M.,
Fisher, R. J., and Kasprzak, K. S. (2003) The octapeptidic end of the
C-terminal tail of histone H2A is cleaved off in cells exposed to
carcinogenic nickel(II). Chem. Res. Toxicol. 16, 1555−1559.
(60) Zavitsanos, K., Nunes, A. M., Malandrinos, G., Kallay, C.,
Sovago, I., Magafa, V., Cordopatis, P., and Hadjiliadis, N. (2008)
Interaction of Cu(II) and Ni(II) with the 63−93 fragment of histone
H2B. Dalton Trans., 6179−6187.
(61) Nunes, A. M., Zavitsanos, K., Del Conte, R., Malandrinos, G.,
and Hadjiliadis, N. (2009) Interaction of histone H2B (fragment 63−
93) with Ni(II). An NMR study. Dalton Trans., 1904−1913.
(62) Nunes, A. M., Zavitsanos, K., Del Conte, R., Malandrinos, G.,
and Hadjiliadis, N. (2010) The possible role of 94−125 peptide
fragment of histone H2B in nickel-induced carcinogenesis. Inorg.
Chem. 49, 5658−5668.
(63) Nunes, A. M., Zavitsanos, K., Malandrinos, G., and Hadjiliadis,
N. (2010) Coordination of Cu2+ and Ni2+ with the histone model
Chemical Research in Toxicology Article
dx.doi.org/10.1021/tx500337w |Chem. Res. Toxicol. 2014, 27, 1996−20092008
peptide of H2B N-terminal tail (1−31 residues): A spectroscopic
study. Dalton Trans. 39, 4369−4381.
(64) Schmidt, M., Raghavan, B., Muller, V., Vogl, T., Fejer, G.,
Tchaptchet, S., Keck, S., Kalis, C., Nielsen, P. J., Galanos, C., Roth, J.,
Skerra, A., Martin, S. F., Freudenberg, M. A., and Goebeler, M. (2010)
Crucial role for human Toll-like receptor 4 in the development of
contact allergy to nickel. Nat. Immunol. 11, 814−819.
(65) Vennegaard, M. T., Dyring-Andersen, B., Skov, L., Nielsen, M.
M., Schmidt, J. D., Bzorek, M., Poulsen, S. S., Thomsen, A. R.,
Woetmann, A., Thyssen, J. P., Johansen, J. D., Odum, N., Menne, T.,
Geisler, C., and Bonefeld, C. M. (2014) Epicutaneous exposure to
nickel induces nickel allergy in mice via a MyD88-dependent and
interleukin-1-dependent pathway. Contact Dermatitis 71, 224−232.
(66) Lizarbe, M. A., Barrasa, J. I., Olmo, N., Gavilanes, J., and Turnay,
J. (2013) Annexin-phospholipid interactions. Functional implications.
Int. J. Mol. Sci. 14, 2652−2683.
(67) Gavins, F. N., and Hickey, M. J. (2012) Annexin A1 and the
regulation of innate and adaptive immunity. Front. Immunol. 3, 354.
(68) Caron, D., Maaroufi, H., Michaud, S., Tanguay, R. M., and
Faure, R. L. (2013) Annexin A1 is regulated by domains cross-talk
through post-translational phosphorylation and SUMOYlation. Cell.
Signalling 25, 1962−1969.
(69) Bharadwaj, A., Bydoun, M., Holloway, R., and Waisman, D.
(2013) Annexin A2 heterotetramer: structure and function. Int. J. Mol.
Sci. 14, 6259−6305.
(70) Madureira, P. A., and Waisman, D. M. (2013) Annexin A2: the
importance of being redox sensitive. Int. J. Mol. Sci. 14, 3568−3594.
(71) Mack, D. P., and Dervan, P. B. (1992) Sequence-specific
oxidative cleavage of DNA by a designed metalloprotein, Ni(II).GGH-
(Hin139−190). Biochemistry 31, 9399−9405.
(72) Liang, R., Senturker, S., Shi, X., Bal, W., Dizdaroglu, M., and
Kasprzak, K. S. (1999) Effects of Ni(II) and Cu(II) on DNA
interaction with the N-terminal sequence of human protamine P2:
enhancement of binding and mediation of oxidative DNA strand
scission and base damage. Carcinogenesis 20, 893−898.
(73) Zavitsanos, K., Nunes, A. M., Malandrinos, G., and Hadjiliadis,
N. (2011) DNA strand breakage induced by CuII and NiII, in the
presence of peptide models of histone H2B. J. Inorg. Biochem. 105,
1329−1337.
(74) Reutelingsperger, C. P., Heerde, W., Hauptmann, R., Maassen,
C., Gool, R. G., Leeuw, P., and Tiebosch, A. (1994) Differential tissue
expression of Annexin VIII in human. FEBS Lett. 349, 120−124.
(75) Lu, L. K., Warshaw, E. M., and Dunnick, C. A. (2009)
Prevention of nickel allergy: the case for regulation? Dermatol. Clin. 27,
155−161.
Chemical Research in Toxicology Article
dx.doi.org/10.1021/tx500337w |Chem. Res. Toxicol. 2014, 27, 1996−20092009
